=head1 NAME
ShiftJIS::Collate - collation of Shift-JIS strings
=head1 SYNOPSIS
use ShiftJIS::Collate;
@sorted = ShiftJIS::Collate->new(%tailoring)->sort(@source);
=head1 DESCRIPTION
This module provides some functions to compare and sort strings
in Shift-JIS based on B<JIS X 4061:1996>,
collation of Japanese character strings, "Nihongo mojiretsu shogo junban".
This module is an implementation of JIS X 4061:1996 and
the collation rules are based on that standard.
See Conformance to the Standard.
=head2 Constructor and Tailoring
The C<new> method returns a collator object.
$Collator = ShiftJIS::Collate->new(
ignoreChar => $regexIgnoredChar,
kanji => $kanji_class,
katakana_before_hiragana => $bool,
level => $collationLevel,
position_in_bytes => $bool,
tounicode => \&sjis_to_unicode,
preprocess => \&preprocess,
upper_before_lower => $bool,
);
# if %tailoring is false (empty),
# $Collator should do the default collation.
=over 4
=item ignoreChar
If specified as a regular expression,
any characters that match it are ignored on collation.
e.g. If you want to ignore C<KATAKANA-HIRAGANA PROLONGED SOUND MARK>
and its halfwidth form, say
ignoreChar => '^(?:\x81\x5B|\xB0)',
=item katakana_before_hiragana
By default, hiragana is before katakana.
If the parameter is true, this is reversed.
=item kanji
Set the kanji class. See L<Kanji Classes>.
Class 1: 'saisho' (minimal)
Class 2: 'kihon' (basic)
Class 3: 'kakucho' (extended)
The kanji class is specified as 1, 2, or 3. If omitted, class 2 is applied.
This module does not support 'kakucho' kanji class,
since the repertoire of Shift-JIS does not define
all the Unicode CJK unified ideographs.
But if the kanji class 3 is specified, you can collate kanji
in the unicode order. In this case you must provide
L<tounicode> coderef which gives a unicode codepoint
from a Shift-JIS character.
=item level
Set the maximum level. See L<Collation Levels>.
Any higher levels than the specified one are ignored.
Level 1: alphabetic ordering
Level 2: diacritic ordering
Level 3: case ordering
Level 4: script ordering
Level 5: width ordering
The collation level is specified as a number
between 1 and 5. If omitted, level 4 is applied.
=item tounicode
If you want to collate kanji in the unicode order,
specify a coderef which gives a unicode codepoint
from a Shift-JIS character.
Such a subroutine should map a string comprising of
a kanji of level 1 and 2 in Shift-JIS to a codepoint
in the range between 0x4E00 and 0x9FFF.
=item position_in_bytes
By default, the C<index> method returns its results in characters.
If this parameter is true, it returns the results in bytes.
=item preprocess
If specified, the coderef is used to preprocess
before the formation of sort keys.
I.e. a sort key is formed from a return value returned by the coderef.
=item upper_before_lower
By default, lowercase is before uppercase.
If the parameter is true, this is reversed.
=back
=head2 Comparison
=over 4
=item C<$result = $Collator-E<gt>cmp($a, $b)>
Returns C<1> (when C<$a> is greater than C<$b>)
or C<0> (when C<$a> is equal to C<$b>)
or C<-1> (when C<$a> is lesser than C<$b>).
=item C<$result = $Collator-E<gt>eq($a, $b)>
=item C<$result = $Collator-E<gt>ne($a, $b)>
=item C<$result = $Collator-E<gt>gt($a, $b)>
=item C<$result = $Collator-E<gt>ge($a, $b)>
=item C<$result = $Collator-E<gt>lt($a, $b)>
=item C<$result = $Collator-E<gt>le($a, $b)>
They works like the same name operators as theirs.
=back
=head2 Sorting
=over 4
=item C<$sortKey = $Collator-E<gt>getSortKey($string)>
Returns a sort key.
You compare the sort keys using a binary comparison
and get the result of the comparison of the strings.
$Collator->getSortKey($a) cmp $Collator->getSortKey($b)
is equivalent to
$Collator->cmp($a, $b)
=item C<@sorted = $Collator-E<gt>sort(@source)>
Sorts a list of strings by B<tanjun shogo>: 'the simple collation'.
=item C<@sorted = $Collator-E<gt>sortYomi(@source)>
Sorts a list of references to arrays of (spell, reading)
by B<yomi-hyoki shogo>: 'the collation using readings and spells'.
B<Yomi-hyoki shogo> is carried out through two comparison stages.
First, order by reading;
next, order by spelling among strings having the same reading.
See also F<sample/yomi.txt>.
=item C<@sorted = $Collator-E<gt>sortDaihyo(@source)>
Sorts a list of references to arrays of (spell, reading)
by B<kan'i-daihyo-yomi shogo>:
'the simplified representative reading collation'.
B<Note>: This module does not support B<kihon-daihyo-yomi shogo>:
'the basic representative reading collation',
in which B<daihyo-yomi jisho>, a dictionary of representative readings,
is required.
B<kan'i-daihyo-yomi shogo> is carried out through B<five> comparison stages.
This ordered list is an example of the result of C<"kan'i-daihyo-yomi shogo">.
(1) Compare the character class of the first character of the spell.
(2) Compare the first character of the reading.
(3) Compare the first character of the spell.
(4) Compare the whole string of the reading.
(5) Compare the whole string of the spell.
See also F<sample/daihyo.txt>.
=back
=head2 Searching
=over 4
=item C<$position = $Collator-E<gt>index($string, $substring)>
=item C<($position, $length) = $Collator-E<gt>index($string, $substring)>
If C<$substring> matches a part of C<$string>, returns
the position of the first occurrence of the matching part in scalar context;
in list context, returns a two-element list of
the position and the length of the matching part.
B<Notice> that the length of the matching part may differ from
the length of C<$substring>.
If C<$substring> does not match any part of C<$string>,
returns C<-1> in scalar context and an empty list in list context.
If your C<substr> function is aware of Shift-JIS,
specify true as C<position_in_bytes>. See L<Constructor and Tailoring>.
=back
=head1 NOTE
=head2 Collation Levels
The following criteria are considered in order
until the collation order is determined.
By default, Levels 1 to 4 are applied and Level 5 is ignored,
for JIS X 4061 does not specify level 5.
=over 4
=item Level 1: alphabetic ordering.
Character classes ordered as following:
1 space
2 'kijutsu kigou' (punctuation marks)
3 'kakko kigou' (quotes, parentheses, and braces)
4 'gakujutsu kigou' (mathematical and technical signs)
5 'ippan kigou' (general symbols)
6 'tan-i kigou' (unit signs)
7 digits (European)
8 'ouji kigou' (Greek and Cyrillic letters, as "European signs")
9 Latin alphabets
10 kana (hiragana and katakana)
11 kanji
12 'geta' mark
In the class, alphabets are collated alphabetically;
kana letters are AIUEO-betically (in the Gozyuon order);
kanji are in the JIS X 0208 order.
According to JIS X 4061, C<GETA MARK> (0x81AC, U+3013)
must be the greatest character (ordered at the last).
Any character for which the order is no defined,
like control characters, box drawings, unassigned characters, etc.
is regarded as a completely ignorable character.
=item Level 2: diacritic ordering.
In kana, the order is as shown the following list.
A voiceless kana, the voiced, then the semi-voiced (if exists).
=item Level 3: case ordering.
A small Latin is lesser than the corresponding Capital.
In kana, the order is as shown the following list.
see Replacement of PROLONGED SOUND MARK and ITERATION MARKs.
Replaced PROLONGED SOUND MARKs (U+30FC and U+FF70);
Small Kana;
Replaced ITERATION MARKs (U+309D, U+309E, U+30FD, and U+30FE);
then, Normal Kana.
=item Level 4: script ordering.
Any hiragana is lesser than the corresponding katakana.
=item Level 5: width ordering.
A character that belongs to the block
C<Halfwidth and Fullwidth Forms>
is greater than the corresponding normal character.
B<BN: JIS X 4061 does not mention this level.
Level 5 is an extention by this module.>
=back
=head2 Kanji Classes
JIS X 4061 specifies three Kanji Classes.
This modules supports 'Saisho' and 'Kihon' Kanji Classes.
=over 4
=item the 'saisho' (minimal) kanji class
It comprises five kanji-like characters:
0x8156 (U+3003)
0x8157 (U+4EDD)
0x8158 (U+3005)
0x8159 (U+3006)
0x815A (U+3007)
Any kanji except U+4EDD are ignored on collation.
=item the 'kihon' (basic) kanji class
It comprises JIS level 1 and 2 kanji in addition to
the minimal kanji class. Sorted in the JIS codepoint order.
Any kanji excepting those defined by JIS X 0208 are ignored on collation.
=item the 'kakucho' (extended) kanji class
All the CJK Unified Ideographs in addition to
the minimal kanji class. Sorted in the Unicode codepoint order.
=back
=head2 Replacement of PROLONGED SOUND MARKs and ITERATION MARKs
SJIS UCS Name
0x815B U+30FC KATAKANA-HIRAGANA PROLONGED SOUND MARK
0xB0 U+FF70 HALFWIDTH KATAKANA-HIRAGANA PROLONGED SOUND MARK
0x8154 U+309D HIRAGANA ITERATION MARK
0x8155 U+309E HIRAGANA VOICED ITERATION MARK
0x8152 U+30FD KATAKANA ITERATION MARK
0x8153 U+30FE KATAKANA VOICED ITERATION MARK
These characters, if replaced, are secondary equal to
the replacing kana, while ternary not equal to.
=over 4
=item KATAKANA-HIRAGANA PROLONGED SOUND MARKs
The PROLONGED MARKs (including the halfwidth equivalent)
are repleced to a normal vowel or nasal
katakana corresponding to the preceding kana if exists.
=item HIRAGANA- and KATAKANA ITERATION MARKs
The ITERATION MARKs (VOICELESS) are repleced
to a normal kana corresponding to the preceding kana if exists.
=item HIRAGANA- and KATAKANA VOICED ITERATION MARKs
The VOICED ITERATION MARKs are repleced to a voiced kana
corresponding to the preceding kana if exists.
=item Cases of no replacement
Otherwise, no replacement occurs. Especially in the
cases when these marks follow any character except kana.
The unreplaced characters are primary greater than any kana.
e.g. CJK Ideograph followed by PROLONGED SOUND MARK
Digit followed by ITERATION MARK
Kana that has no voiced variant followed by VOICED ITERATION MARK
=back
=head2 Conformance to the Standard [JIS X 4061, 6.2]
(1) charset: Shift-JIS.
(2) No limit of the number of characters in the string considered
to collate.
(3) No character class is added.
(4) The following characters are added as collation elements.
IDEOGRAPHIC SPACE is added in the space class.
ACUTE ACCENT, GRAVE ACCENT, DIAERESIS, CIRCUMFLEX ACCENT
are added in 'kijutsu kigou'.
APOSTROPHE, QUOTATION MARK are added in 'kakko kigou'.
HYPHEN-MINUS is added in 'gakujutsu kigou'.
(5) Collation of Latin alphabets with macron and with circumflex
is not supported.
(6) Kanji Classes:
i) the minimal kanji class (Five kanji-like chars).
ii) the basic kanji class (Levels 1 and 2 kanji of JIS)..
=head1 AUTHOR
SADAHIRO Tomoyuki <SADAHIRO@cpan.org>
Copyright(C) 2001-2007, SADAHIRO Tomoyuki. Japan. All rights reserved.
This module is free software; you can redistribute it
and/or modify it under the same terms as Perl itself.
=head1 SEE ALSO
=over 4
=item JIS X 4061
Collation of Japanese character strings
=item JIS X 0201
7-bit and 8-bit coded character sets for information interchange
=item JIS X 0208
7-bit and 8-bit double byte coded KANJI sets for information interchange
=item JIS X 0221
Information technology - Universal Multiple-Octet Coded
Character Set (UCS) - part 1 : Architectute and Basic Multilingual Plane
=item Japanese Industrial Standards Committee (JISC)
L<http://www.jisc.go.jp/>
=item Japanese Standards Association (JSA)
L<http://www.jsa.or.jp/>
=item L<ShiftJIS::String>
=item L<ShiftJIS::Regexp>
=back
=cut